Fig 2: Effects of TTFields on TMZ + CCNU induced DNA damage and on the FA-BRCA DNA repair pathway in GBM cells. A MGMT-expressing U-118 MG and LN-18 cells, and MGMT-non-expressing U-87 MG and LN-229 cells were treated for 72 h with TMZ + CCNU, TTFields (200 kHz, 0.83 V/cm RMS), or both treatments together, followed by immunofluorescence staining of ɣH2AX (green) for detection of DNA damage and DAPI (blue) for nuclear visualization. Presented are representative images at × 20 magnification, and quantification of mean foci per nucleus. Chemotherapy doses: 30 µM TMZ and 10 µM CCNU for U-118 MG cells; 100 µM TMZ and 15 µM CCNU for LN-18 cells; 25 µM TMZ and 8.7 µM CCNU for U-87 MG cells; and 12 µM TMZ and 5 µM CCNU for LN-229 cells. Values are mean ± SEM. *p < 0.05 and **p < 0.01 relative to TTFields + TMZ + CCNU; Dunnett’s multiple comparisons. B Cells were treated for 72 h with TTFields (200 kHz, 0.83 V/cm RMS), followed by immunoblotting of cell lysates for expression of FANCB, FANCD2, FANCJ, and BRCA2. Values are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 relative to control; 2-way ANOVA, Sidak’s multiple comparison

Fig 3: Transcriptomic changes associated with E7820.a Transcriptomic changes associated with E7820. RNA-seq was performed in six cell lines to compare the changes in gene expression induced by E7820. The expression changes (log fold change; log FC) were highly concordant between DLD1-P and DLD1-KO cells (r = 0.95, left panel) or PC9 cells (r = 0.86, middle panel). The heatmap on the right shows that the trends in expression profile changes associated with E7820 were similar between any pair of DLD1-P, DLD1-KO, PC9, FUJI, H2228, and KMLS1. b The genes that were commonly upregulated (fold change > 1.1) and downregulated (fold change < 0.9) by E7820 treatment among the six cell lines included 1655 and 2787 genes, respectively. Ingenuity pathway analysis (IPA) was conducted for pathway analysis. Among the identified pathways shared by upregulated genes, “Protein Ubiquitination Pathway” and “Spliceosomal Cycles” were highly ranked (−log(p value) = 9.01 and 3.33, respectively). In contrast, the enriched pathways of the downregulated genes included pathways related to DNA damage repair, such as “Nucleotide excision repair (NER) Enhanced Pathway”, “Hereditary Breast Cancer Signaling”, and “Role of BRCA1 in DNA Damage Response” (−log(p value) = 9.27, 8.51 and 7.38, respectively). c The changes in gene expression induced by E7820 treatment are shown. The fold changes are shown in a heatmap comparing the expression of cell lines after E7820 treatment (1 μM for 48 h) to basal expression. The genes are involved in homologous recombination repair (HRR), nonhomologous end joining (NHEJ), the Fanconi anemia complementation group (FANC) pathway, nucleotide excision repair (NER), base excision repair (BER) and other pathways related to DNA damage repair (Common). Only genes whose expression was decreased by 10%< are shown. Information on all genes is shown in Supplemental Fig. 4. d The number of splicing anomalies induced by E7820 treatment is indicated. Similar alternative splicing events were induced by E7820 in both DLD1-P and DLD1-KO cells. Mis-splicing event codes are as follows: TSS transcription start site, A5SS alternative 5’ splice site, A3SS alternative 3’ splice site, SES single-exon skipping, MES multiple-exon skipping, IR intron retention, OR overlapping region. e The changes in IR ratio induced by E7820 treatment are shown. The most significant changes in individual genes of Fig. 3c are shown in a heatmap comparing the IR ratios of cell lines after E7820 treatment (1 μM for 48 h) to basal ratio. f Full-length cDNAs were selectively prepared from the cells and sequenced with a Sequel II long-read sequencer (Pacific Biosystems). The long-read sequences were used to identify isoforms using the bulk Iso-seq workflow and pigeon workflow. The isoforms were visualized using IGV (Integrative Genomics Viewer). The individual long reads were tagged with PB ID, as shown on the left. Intron retention within FANCD2 was specifically observed in the cells treated with E7820 (arrowheads). g The protein levels of FANCD2 and FANCA in DLD1-KO cells. DLD1-KO cells were incubated with E7820 for the indicated times, and western blotting was conducted to evaluate the protein expression of FANCD2, FANCA, γH2AX, and cleaved caspase-3.

Fig 4: ERK activation reverses the tumour‐suppressive effect of BA in cisplatin‐treated xenografts. (A) Representative images of xenograft tumours and quantitative analysis of tumour volumes in four groups: Control (NC), cisplatin (CDDP), BA + CDDP, and BA + CDDP + ERK agonist (ERKA). BA + CDDP treatment markedly reduces tumour growth compared with CDDP alone, while ERK activation restores tumour size to near‐control levels. Data are shown as mean ± SEM; *p < 0.05 vs. NC, #p < 0.05 vs. CDDP, †p < 0.05 vs. BA + CDDP. (B) Western blot analysis of xenograft lysates showing decreased UBE2T as well as monoubiquitination of FANCI and FANCD2 in the BA + CDDP group, along with suppressed ERK1/2 phosphorylation; ERK agonist treatment restores p‐ERK1/2 and UBE2T expression. β‐Actin served as the loading control for total protein. (C, D) Immunohistochemical staining of xenograft tumour sections confirming reduced UBE2T (C) and p‐ERK1/2 (D) expression in the BA + CDDP group, with re‐elevation following ERK agonist co‐treatment. Scale bar = 20 μm.

Fig 5: Ba suppresses FANCI/FANCD2 monoubiquitination, FANCD2 foci formation, and downstream DNA repair complex assembly in glioma cells. (A–C) Western blot analysis of FANCI and FANCD2 monoubiquitination in U87MG (A), LN229 (B), and U251 (C) cells treated with cisplatin (CDDP) or BA plus CDDP. Cisplatin induces robust monoubiquitination of both FANCI and FANCD2, whereas BA pretreatment markedly diminishes these modifications (red arrows), indicating inhibition of FA pathway activation. β‐Actin served as a loading control. (D–F) Immunofluorescence analysis show the FANCD2 nuclear foci formation in U87MG (D), LN229 (E), and U251 (F) cells. Blue: DAPI‐stained nuclei; green: FANCD2 Foci. Quantification (right) revealed a significant decrease in the percentage of FANCD2‐positive nuclei in BA + CDDP‐treated cells compared with CDDP alone. Data are presented as mean ± SEM (n = 3 independent experiments); *p < 0.05 vs. NC, #p < 0.05 vs. CDDP. (G–I) Co‐immunoprecipitation (Co‐IP) assays examining FANCD2‐associated repair proteins. In all three glioma cell lines, cisplatin enhances the interaction between FANCD2 and REV1, BRCA1 and ERCC1, reflecting activation of translesion synthesis (TLS) and homologous recombination (HR) branches of DNA repair. BA pretreatment substantially weakens these interactions, suggesting that BA not only blocks FANCD2 monoubiquitination but also disrupts the recruitment of downstream repair effectors.